Generative AI Engineer

About the role

Magpie Health Analytics is looking for an experienced Generative AI Engineer to design, build, and operate secure AI solutions for government health payers. The role focuses on creating a decision‑support system that lets users search large bodies of program documentation using natural language.

Key responsibilities

• Design, develop, and maintain generative AI solutions that ground large language model responses in authoritative source documentation.
• Build and maintain document ingestion pipelines, including parsing, chunking, metadata tagging, embedding generation, and vector/semantic search indexing.
• Collaborate with project managers, data engineers, full‑stack developers, analysts, and client stakeholders to translate user needs into prompts, evaluation criteria, and system requirements.
• Implement source citations, guardrails, content filtering, PII/PHI protections, and refusal handling for out‑of‑scope queries.
• Develop evaluation frameworks and test sets to measure retrieval quality, answer accuracy, groundedness, and consistency.
• Deploy and operate solutions on AWS services while supporting federal security controls and AI governance reviews.

Required profile

• Master’s degree in computer science, data science, engineering, or a related field.
• 5+ years of software and cloud‑based development experience, with recent work on LLM or generative AI applications.
• Strong problem‑solving and communication skills, able to explain AI capabilities to non‑technical stakeholders.
• Experience working independently and within cross‑functional teams.

Required skills

• Python programming.
• Modern LLM APIs and SDKs (e.g., Hugging Face, LangChain, LlamaIndex).
• AWS AI and data services such as Amazon Bedrock, OpenSearch, Kendra, SageMaker, Lambda, and S3.
• Prompt engineering, context engineering, and techniques to reduce hallucinations.
• Vector databases, embeddings, and hybrid/semantic search.
• CI/CD pipelines and infrastructure as code tools (GitHub Actions, CloudFormation, Terraform, Jenkins).